///2014 Abstract Details
2014 Abstract Details2019-07-18T14:34:47+00:00

Entering paper-based medical records into a simple, low-cost database in a low resource country allows for more effective tracking of quality and outcomes measures.

Abstract Number: T-26
Abstract Type: Original Research

William D Nelson MD1 ; Srofenyoh Emanuel MD2; Petermann Cynthia --3; Haruna Rabiatu --4; Adeyemi J Olufolabi MB BS5

Background: Ridge Regional Hospital in Accra, Ghana, entered into collaboration with Kybele (Kybele, Inc., Winston-Salem, NC), in 2007 to advance the care of pregnant women and newborns. The hospital uses a paper-based medical record. In 2011, we designed and developed a low-cost database (MS Access, Microsoft Corporation, Redmond, WA) to store specific data points related to the pregnancy, delivery and neonatal outcome of each patient. This database was designed to work alongside the paper-based health records at Ridge Hospital, not as a replacement. After delivery, a patient's chart is forwarded to a medical records employee who transcribes approximately 50 data points into the database. The aim of this project was to validate the accuracy of data entry into the electronic database performed by a single individual.

Methods: We selected 6 data points that we considered important variables to follow, each of which are collected at the time of admission or immediately following delivery: Age, blood pressure, presence of proteinuria, fetal heart rate, delivery method (vaginal delivery vs. cesarean), and Apgar scores (1 and 5 minutes).

We determined the number of charts necessary to sample 5% of deliveries over a 3 month period, and then randomly selected the necessary number of charts. Each paper chart was reviewed by a medical records employee who was blinded to the database data. This employee read aloud each of the six data points of interest, while a researcher (WN) viewed the corresponding database data. Each data point was evaluated individually, and inconsistencies between the paper charts and the database records were noted.

Results: A total of 1511 deliveries occurred from October 1, 2013 through December 31, 2013. We reviewed 5% (76) of these charts and evaluated 6 data points per chart, for a total of 456 data points. There were 11 discrepancies noted during the comparison of paper charts and database records, yielding an error rate of 2.41%. (See attached table.)

Conclusion: This retrospective sampling suggests that paper-based medical information can be entered into a low-cost electronic database by a dedicated employee with a high degree of accuracy. Converting key obstetric patient data from a paper-based format to a simple database, although labor intensive, allows non-technical end-users to easily generate clinical outcome and quality measure reports on a more frequent basis than would otherwise be possible with only paper-based data.

SOAP 2014